B+ tree selection index (1) -- MySQL from entry to proficiency (22)

The last article introduced that grouping with index columns can also improve efficiency , And the cost of returning to the table , The back table is a secondary index + Return to the table for query , If the amount of data returned is too large ,mysql The optimizer will use full table scan queries , The overlay index will not be returned to the table .

B+ Tree index uses (9) grouping 、 Back to the table 、 Overlay index （ The 21st )

How to select indexes

After understanding the power of index , How do we use indexes ？

search in progress 、 Sort 、 Create indexes for grouped Columns

That is to say where Column after statement , Join column in join clause , Or appear in order by 、group by The columns in the clause need to be indexed , and select Query columns do not need . such as

Select birthday,country from person_info where name = ‘’; This sql Statement inside birthday and country There's no need to create indexes , and name You need to create an index .

Consider the cardinality of the column

The cardinality of a column refers to the non duplicate data in a column , such as name Listed 3,2,3,5,5,3,2,3,3, Although there are many data , But his different data is only three , So the base number is 3. In the data record , The larger the cardinality of the column , The more scattered the data , The smaller the cardinality of the column , It shows that the more centralized the data . Suppose the cardinality of the column is 1, All the data in the front row are the same , You can't sort , There is no point in indexing , If you build a secondary index , It will also lose performance due to the operation of returning a large amount of data , So only when the column cardinality is larger , Is suitable for indexing .

The type of index column should be as small as possible

We are defining the specified column types displayed in the table structure , Let's take integer type as an example , Yes TINYINT,MEDIUMINT,INT,BIGINT How many , The memory they occupy increases in turn , The type size we refer to here , Is to establish the data range of this column , Minimize choices , This is because ：1） The smaller the data type , The faster the query （CPU Level things ）.2） The smaller the amount of data , The smaller the storage space occupied by the index , Put more records on a data page , This reduces the number of disks I/O Performance loss , Reduce the performance consumption caused by querying more data pages .

This is not only applicable to secondary indexes , The same applies to primary keys , Because the leaf nodes of the secondary index have primary keys , The smaller the type of primary key , The smaller the memory used .

Index string prefix

We know that a string consists of several characters , We use it utf8 The composition needs to occupy 1~3 Bytes ,1) Spatiality ：b+ The tree needs to store all the strings of the columns , The longer the string ,b+ The larger the space occupied by the leaf node of the tree .2） Timeliness ： The longer the string , When sorting and comparing , The longer it takes to compare one character to another .

We talked about that before , For columns of strings , It is sorted by character , So the index designer , Considering space and time , Only the first few specified characters of the string are stored and sorted , In this way, you can also relatively locate your seat , Using the corresponding primary key id De clustering index b+ The tree queries all values . Write sql We can write this when , establish name by 100 String range of , But the index specification only specifies name front 10 Characters of .

mysql> create table person_info2(
    -> id int not null auto_increment,
    -> name varchar(100) not null,
    -> birthday date not null,
    -> phone char(11) not null,
    -> country varchar(100) not null,
    -> primary key (id),
    -> key idx_name_birthday_phone (name(10),birthday,phone)
    -> );
Query OK, 0 rows affected (0.04 sec)

// You can also create long string columns hash Indexes ：
mysql> create index hash_name using hash on person_info2(name);
Query OK, 0 rows affected (0.04 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> show index from person_info2;
+--------------+------------+-------------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table        | Non_unique | Key_name                | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+--------------+------------+-------------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| person_info2 |          0 | PRIMARY                 |            1 | id          | A         |           0 |     NULL | NULL   |      | BTREE      |         |               |
| person_info2 |          1 | idx_name_birthday_phone |            1 | name        | A         |           0 |       10 | NULL   |      | BTREE      |         |               |
| person_info2 |          1 | idx_name_birthday_phone |            2 | birthday    | A         |           0 |     NULL | NULL   |      | BTREE      |         |               |
| person_info2 |          1 | idx_name_birthday_phone |            3 | phone       | A         |           0 |     NULL | NULL   |      | BTREE      |         |               |
| person_info2 |          1 | hash_name               |            1 | name        | A         |           0 |     NULL | NULL   |      | BTREE      |         |               |
+--------------+------------+-------------------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+

Copy